HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Renné, T. Articles by Müller-Esterl, W. Search for Related Content PubMed PubMed Citation Articles by Renné, T. Articles by Müller-Esterl, W. Social Bookmarking                      What's this?

J Biol Chem, Vol. 274, Issue 36, 25777-25784, September 3, 1999

Mapping of the Discontinuous H-kininogen Binding Site of Plasma Prekallikrein
EVIDENCE FOR A CRITICAL ROLE OF APPLE DOMAIN-2^*

Thomas Renné, Jürgen Dedio, Joost C. M. Meijers^§, Dominic Chung^¶, and Werner Müller-Esterl

From the Institute of Physiological Chemistry and Pathobiochemistry, Johannes Gutenberg University at Mainz, Duesbergweg 6, D-55099 Mainz, Germany, the  Department of Haematology, University Medical Center Utrecht, Heidelberglaan 100, NL-3584 CX Utrecht, The Netherlands, and the ^¶ Department of Biochemistry, University of Washington, Seattle, Washington 98195

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Plasma prekallikrein, a zymogen of the contact phase system, circulates in plasma as heterodimeric complex with H-kininogen. The binding is mediated by the prekallikrein heavy chain consisting of four apple domains, A1 to A4, to which H-kininogen binds with high specificity and affinity (K_D = 1.2 × 10⁸ M). Previous work had demonstrated that a discontinuous kininogen-binding site is formed by a proximal part located in A1, a distal part exposed by A4, and other yet unidentified portion(s) of the kallikrein heavy chain. To detect relevant binding segment(s) we recombinantly expressed single apple domains and found a rank order of binding affinity for kininogen of A2 > A4  A1 > A3. Removal of single apple domains in prekallikrein deletion mutants reduced kininogen binding by 21 (A1), 64 (A2), and 24% (A4), respectively, whereas deletion of A3 was without effect. Transposition of homologous A2 domain from prekallikrein to factor XI conferred high-affinity kininogen binding from the former to the latter. The principal role of A2 for H-kininogen docking to the prekallikrein heavy chain was further substantiated by the finding that cleavage of a single peptide bond in A2 drastically diminished the H-kininogen binding affinity. Furthermore, the epitope of monoclonal antibody PKH6 which blocks kallikrein-kininogen complex formation with an IC₅₀ of 8 nM mapped to the center portion of domain A2. Our data indicate that domain A2 and two flanking sequence segments of A1 and A4 form a discontinuous binding platform for H-kininogen on the prekallikrein heavy chain. Domain-specific antibodies directed to these critical sites efficiently interfered with contact phase-induced bradykinin release from H-kininogen.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Human plasma prekallikrein (PPK),¹ the zymogen of the plasma serine proteinase -kallikrein is involved in the intrinsic pathway of blood coagulation (1, 2) in pro-urokinase-dependent fibrinolysis (3, 4), and in local inflammation (5). The zymogen is converted into its active form by surface-bound activated factor XII (FXIIa) (6, 7) via cleavage of a single peptide bond at position 371. The active enzyme, -kallikrein, is composed of a catalytically active light chain of 35 kDa and a heavy chain of 50 kDa, linked together by a single disulfide bridge (8). Autocatalytic cleavage at Lys¹⁴⁰-Ala¹⁴¹ of its heavy chain further converts -kallikrein into a three-chain form, -kallikrein (9). Analyses revealed that the 371 residues of the PPK heavy chain is composed of four tandem repeats of 90-91 amino acid residues each (10) with a unique disulfide bridge pattern where the first and sixth, second and fifth, and third and forth cysteine residues are linked (11). These repetitive modules, aptly dubbed "apple" domains A1 to A4, mediate the high-affinity binding of PPK to its major substrate, high molecular mass kininogen (H-kininogen, HK), with an apparent K_D of 1.2 × 10⁸ M (12). The bimolecular complex docks to the plasma membranes of many cells via specific and affine cell-binding sites exposed on HK domains D3 and D5_H. Local accumulation of the prohormone and its cognate processing enzyme on the surface of target cells such as neutrophils, platelets, and endothelial cells allows the extremely short-lived effector of the system, bradykinin (BK; t_1/2 < 15 s), to act on cellular receptors next to the site of release (13).

To meet the requirements of a locally operating effector system, an elaborate network of complementary structures ensures that HK and PPK interact both in solution and on surfaces. HK exposes a continuous segment of 27 amino acids in the carboxyl-terminal portion of domain D6_H of its light chain to which PPK binds (12, 14) whereas the corresponding HK-binding site on the prekallikrein heavy chain is highly discontinuous. Affinity cross-linking studies indicated that one interacting segment is localized in the amino-terminal portion of A1 (15). An antibody-based strategy as well as peptide competition studies identified a second binding segment in the center part of A4 (16-18) and indicated that other, yet unknown portion(s) of the PPK heavy chain contribute to H-kininogen binding (16). On the amino acid level PPK exhibits 58% sequence identity to FXI, another serine proteinase of the contact activation system (1, 19, 20). Like PPK, FXI complexes with HK via its heavy chain though with lower affinity (K_D = 1.8 × 10⁸ M) than PPK (21). Since the FXI-binding site on HK overlaps with the PPK-binding site, the two zymogens mutually displace each other from the HK light chain (14, 21).

The aim of the present study was the identification of crucial structures and domains that make up the discontinuous HK-binding site in PPK. By direct binding studies with recombinantly expressed single apple domains, by analysis of deletion mutants and chimeras of PPK and FXI where apple domains had been removed and exchanged, and by antibody competition experiments we provide convincing evidence that apple domain A2 in PPK is crucial to HK binding. We demonstrate that blockage of the relevant subsites in domains A1, A2, and A4 efficiently attenuates contact phase-dependent bradykinin release from the HK-PPK complex.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Proteins and Antibodies-- HK and PPK were isolated from human plasma according to established protocols (22, 23). Human FXII was from Enzyme Research Laboratories (South Bend, IN) and activated by incubation with glass beads for 30 min at 37 °C. The generation and characterization of the mouse monoclonal antibodies to human PPK, i.e. PKH1, PKH4, PKH6, and PKL16, have been previously detailed (22). Monoclonal antibody PKH19 was raised against the synthetic peptide PK31 of the human PPK sequence (15). Antisera AS176 and AS199 were raised in rabbits against purified human PPK and FXI, respectively. Recombinantly expressed human PPK apple domain A3 fused to human tissue-type plasminogen activator (tPA; see below) was used to generate antibodies ("anti-rA3") in mice following standard immunization protocols. Monoclonal antibodies HKL16 and HKL22 directed to domain D6_H, HKH14 directed to D3, and MBK3 directed to the kinin moiety of D4 of human HK were used (24). To produce active forms of kallikrein, PPK was incubated with FXIIa at a molar ratio of 100:1 in PBS (6.5 mM Na₂HPO₄, 1.5 mM KH₂PO₄, 2.7 mM KCl, 150 mM NaCl, pH 7.4) at 37 °C for 2 h to give -kallikrein or for 72 h to yield -kallikrein. For biotinylation, 100 µg of HK was incubated with 10 µg of biotin--aminocaproyl-N-hydroxysuccinimide (biotin-X-NHS, Pierce, St. Augustin, Germany) in 0.1 M NaHCO₃ for 4 h at 4 °C. The buffer was changed to 150 mM NaCl, 100 mM NaH₂PO₄, 10 mM Na₂HPO₄, pH 7.4, and unreacted biotin-X-NHS was removed by 3 centrifugations at 2,000 × g at 4 °C using a Microcon-10 column (Amicon, Beverly, MA) with a 10,000 Da cut-off.

Cell Culture-- Human embryonic kidney cells (HEK293t) were cultured under standard conditions in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) containing 4.5 g/liter glucose, 10% (v/v) fetal bovine serum, 0.5% (w/v) penicillin/streptomycin in a humidified CO₂ atmosphere at 37 °C. Baby hamster kidney (BHK) cells were kept under the same conditions except that Dulbecco's modified Eagle's medium was supplemented with fetal bovine serum (5% v/v), penicillin (50 µg/ml), streptomycin (50 µg/ml), and neomycin (100 µg/ml).

Gel Electrophoresis and Western Blotting-- Proteins were resolved by polyacrylamide gel electrophoresis in the presence of 0.1% (w/v) sodium dodecyl sulfate (SDS-PAGE) at 30 mA for 90 min. Marker proteins (Amersham Pharmacia Biotech, Uppsala, Sweden) were phosphorylase b (94 kDa), bovine serum albumin (67 kDa), ovalbumin (43 kDa), and carbonic anhydrase (30 kDa). The resolved proteins were visualized by silver staining or transferred to nitrocellulose at 100 mA for 30 min by the semi-dry technique. The membranes were blocked with PBS containing 5% (w/v) dry milk powder and 0.05% (v/v) Tween 20. For immunoprinting of the transferred proteins (25) the first antibody was applied typically at 1:1,000 dilution in PBS/milk. Bound antibody was detected by a horseradish peroxidase-coupled secondary antibody against mouse immunoglobulin (DAKO, Hamburg, Germany), followed by the chemiluminescence detection system (Amersham).

Binding Assays-- A direct binding assay was employed to analyze HK binding to the various forms of kallikrein. A serial dilution (2ⁿ; starting concentration 8.8 µg/ml = 100 nM) of PPK, -kallikrein, or -kallikrein in 100 mM sodium acetate, 100 mM NaCl, pH 5.5 (coating buffer), was applied to microtiter plates (MaxiSorp, Nunc, Wiesbaden, Germany). The plates were washed 6 times with PBS and blocked with 1% (w/v) bovine serum albumin in PBS for 45 min, followed by incubation with 1 µg/ml (8.3 nM) biotinylated HK in PBS, 1% bovine serum albumin in the presence of protease inhibitor mixture of 10 µg/ml each of soybean trypsin inhibitor, aprotinin, phenylmethylsulfonyl fluoride (Sigma, Deisenhofen, Germany), and 0.1 mM Pefabloc SC (Roth, Karlsruhe, Germany) for 45 min at 37 °C. After washing with PBS bound HK was detected by streptavidin-peroxidase (1 µg/ml; Roche Molecular Biochemicals, Germany) for 45 min, followed by the substrate, 0.15% (w/v) diammonium 2,2'-azido-bis-(3-ethyl-2,3-dihydrobenzthiazoline-6-sulfonate) and 0.012% (v/v) H₂O₂ in 100 mM citric acid, pH 4.5, for 30 min. The change in absorbance was monitored at 405 nm by an enzyme-linked immunosorbent assay (ELISA) reader (Dynatech, Deppendorf, Germany). Alternatively PPK deletion mutants or PPK/FXI chimera (see below) present in the supernatants of transfected cells were used for coating and probed as above. To test the interaction of single recombinant apple domains with HK, fusion proteins were coated in a serial dilution (2ⁿ) starting from 10 µg/ml (200 nM), followed by incubation with 4 µg/ml (33 nM) HK, 2 µg/ml (13.3 nM) of the monoclonal antibody HKH14, and a horseradish peroxidase-coupled secondary antibody. All incubation steps were done at 37 °C for 45 min except for the coating which was done at 4 °C overnight. A sandwich ELISA was used to measure the concentrations of recombinant proteins present in the supernatants of transfected cells. Antibody AS176 to PPK at 1:1000 dilution was used for coating, free binding sites were blocked with PBS, 1% bovine serum albumin, serial dilutions (2ⁿ) of the supernatants were applied, and bound PPK was detected by antibody PKL16 to the PPK light chain followed by a horseradish peroxidase-coupled secondary antibody against mouse immunoglobulin and the chromogenic substrate. A competitive ELISA was established to test for the interference of unlabeled antibodies with HK·PPK complex formation. PPK was coated at 4 µg/ml (45.4 nM) and serial dilutions (2ⁿ, starting concentration 180 µg/ml = 1.2 µM) of antibodies to PPK including 1 µg/ml (8.3 nM) biotinylated HK were applied. Bound biotinylated HK was probed with streptavidin-peroxidase followed by the chromogenic substrate. Values of IC₅₀ were calculated by the KaleidaGraph 3.05 algorithm (Synergy Software, Reading, PA).

Recombinant Expression of PPK Domains in Escherichia coli-- The pMAL-c2 expression system (New England Biolabs, Bad Schwalbach, Germany) was used for expression and purification of fusion proteins consisting of the bacterial maltose-binding protein (MBP) and PPK apple domains in E. coli strain BL21. Polymerase chain reaction (PCR) with Taq polymerase (Amersham Pharmacia Biotech) generated cDNA fragments encoding single PPK apple domains with the following upstream and downstream primers (Roth): 5'-gtttctagaatgattttattcaagcaagc-3' and 5'-cgatggcaaaagcttatttaatgacc-3' (for domain A1), 5'-catcaaataagtgaattccatcgagac-3' and 5'-catgtgggatccaatttcttaaagg-3' (A2), 5'-gccctttctagaattggttgcc-3' and 5'-gcaagcttcaggttaagttcttttgcag-3' (A3) and 5'-cctctagatatagccttttaacctgc-3' and 5'-gcaagcttgtttatgttgtgcagacagag-3' (A4), respectively. Plasmid pPK was used as the template in a polymerase chain reaction that comprises 40 cycles of denaturation at 95 °C for 45 s, annealing at 50 °C for 45 s, and extension at 72 °C for 90 s in a thermal cycler (Biometra, Göttingen, Germany). Before ligating the constructs into the pMAL-c2 vector using T4 DNA-ligase (New England Biolabs), the vector and the isolated PCR products were cleaved with restriction enzymes HindIII and XbaI (A1), EcoRI and BamHI (A2), XbaI and HindIII (A3 and A4), respectively, and purified by phenol-chloroform extraction. Recombinant plasmids were propagated in E. coli XL1-blue strain; vectors were isolated by a plasmid DNA isolation kit (Qiagen, Hilden, Germany) before transfection into E. coli BL21 strain for expression. Exponentially growing cultures containing the relevant constructs were stimulated for 2 h with 0.5 mM isopropyl--D-thiogalactopyranoside (Roth), and the cells were harvested by centrifugation at 4,000 × g for 20 min at 4 °C. The pelleted cells were resuspended in 2 times PBS supplemented with a protease inhibitor mixture (10 µg/ml each soybean trypsin inhibitor, benzamidine, leupeptin (Sigma), and 0.1 mM Pefabloc SC), put on ice, and lysed by repeated brief ultrasonic pulses for 3 min. Following centrifugation at 20,000 × g for 15 min at 4 °C to remove the cell debris, the supernatants containing the MBP-PPK apple fusion proteins were applied to an amylose resin. After extensive washing with PBS, bound proteins were eluted with PBS, 20 mM maltose, followed by gel filtration over a Sephadex 200 column (Amersham Pharmacia Biotech) in PBS. Pooled fractions of the fusion proteins were characterized by SDS-PAGE and Western blotting using domain-specific antibodies.

Expression of Single Apple Domains Fused to tPA-- The cDNA encoding single PPK apple domains, A1 (Gly¹-Ser⁹⁰), A2 (Ala⁹¹-Ile¹⁸⁰), A3 (Gly¹⁸¹-Glu²⁷¹), and A4 (Pro²⁷²-Ser³⁶²) were amplified by Taq polymerase PCR with primers introducing a BglII site and a XhoI site at the 5'- and 3'-ends, respectively, of the amplified DNA. PCR products were cloned into the TA-cloning vector pCRII (Invitrogen, Leek, The Netherlands), and subjected to dideoxy sequencing. The cDNAs were excised from pCRII by BglII and XhoI digestion and directionally cloned into the corresponding sites of the tPA expression vector ZpL7(Ser⁴⁷⁸-Ala) modified as described (16, 26). The corresponding constructs encode fusion proteins with the prepro sequences of tPA, followed by a single PPK apple domain each, kringles 1/2, and the active site-mutated (Ser⁴⁷⁸-Ala) catalytic domain of tPA. The expression plasmids were transfected into BHK cells, and the corresponding fusion proteins expressed in serum-free medium (Opti-MEM, Life Technologies, Inc.) were purified by immunoabsorption using a monoclonal antibody to tPA, as described (16).

Expression of PPK Deletion Mutants and PPK-FXI Chimeras-- To clone the PPK cDNA as an EcoRI unit, an internal EcoRI site encompassing codon GAA for Glu⁴⁷⁸ was eliminated by mutation to GAG using overlap extension with PCR (27) in the plasmid vector pZEM229R (a gift from Donald Foster, ZymoGenetics, Inc.). In these studies, the 5'- and 3'-noncoding sequences in the native PPK cDNA were also removed, and EcoRI sites were introduced immediately before the initiator codon and after the stop codon. In the construction process, a silent mutation in the codon for Thr⁵⁷⁷ (ACA to ACC) was introduced by Taq polymerase. This construct expressing wild-type PPK is designated pZEM-PPK. Gene splicing by overlap extension with PCR (28) was used to construct deletions of each of the four apple domains; the corresponding constructs were designated pZEM-1, pZEM-2, pZEM-3, and pZEM-4, respectively, and the corresponding PPK deletion variants are dubbed 1 to 4. Similarly, chimeric constructs containing parts of FXI and PPK were spliced together by the overlap extension method. Thus, ZEM-XI codes for wild-type factor XI, whereas pZEM7.1 encodes chimera 7.1 in which the signal peptide and heavy chain of PPK are fused to the light chain of FXI, while pZEM7.2 encodes the complementary chimera 7.2 comprising the signal peptide and heavy chain of FXI fused to the light chain of PPK. pZEM6.3 encodes chimera 6.3 in which apple domain A2 in FXI is replaced by A2 of PPK, whereas pZEM6.2 encodes the complementary chimera 6.2 where A2 in PPK is exchanged for A2 of FXI. The sequence of all constructs was verified by dideoxy sequencing. When indicated, appropriate restriction fragments containing the normal sequence were used to replace and correct for misincorporations introduced by Taq polymerase. In expression studies, the respective expression units were excised by EcoRI and cloned into the EcoRI site of pcDNA3(+) (Invitrogen). Orientation of the constructs was confirmed by restriction analysis and the constructs were transfected into HEK293t cells using LipofectAMINE (Life Technologies). The transfection efficiency monitored by parallel transfections with a vector encoding green fluorescent protein (29) was 40%. Fusion proteins were expressed in serum-free medium.

Protein Quantification-- Protein concentrations in supernatants of transfected HEK293t cells were determined by sandwich ELISA (see above) and biospecific interaction analysis using surface plasmon resonance spectroscopy (BIAcore, Freiburg, Germany). CM5 sensor chips were coated with antibodies to PPK (AS176) or FXI (AS199) using the Amine Coupling Kit provided by the manufacturer. Serial dilutions (2ⁿ) of the supernatants were applied at a continuous flow rate of 20 µl/min, and antigen-antibody association was followed for 90 s. Dissociation of the immune complex was induced by applying PBS, and monitored over 3 min. The chip was reconstituted by a brief wash with 30 mM HCl. For calibration supernatants from control cells that had been transfected with an irrelevant vector construct were spiked with varying concentrations of purified PPK. The relative protein concentrations were calculated with the BIAevaluation 2.1 program (BIAcore).

Immunoprecipitation of ³⁵S-Labeled PPK-FXI Chimeras-- HEK293t cells were transfected by the LipofectAMINE method with pcDNA3(+) vectors encoding wild-type FXI, wild-type PPK, or the chimeric constructs 6.2, 6.3, 7.1, and 7.2. After 60 h the cells were washed with Cys/Met-free medium (Dulbecco's modified Eagle's medium) and incubated for 45 min at 37 °C. Cells were labeled with 100 µCi/ml [³⁵S]Cys/Met (Tran³⁵S-label, ICN, Eschwege, Germany) for 12 h at 37 °C, washed 3 times with PBS, and lysed in RIPA (150 mM NaCl, 50 mM Tris-HCl, pH 8.0, 0.1% (m/v) SDS, 0.5% (m/v) deoxycholic acid, 1% (v/v) Nonidet P-40, 10 µg/ml phenylmethylsulfonyl fluoride, 10 µg/ml benzamidine-HCl) under rotation for 1 h at 4 °C. The cell lysate was centrifuged at 14,000 × g for 15 min at 4 °C, the supernatant was transferred to a fresh tube and incubated for 60 min at 4 °C with 20 µl of a mixture of antibodies to PPK (AS176) and FXI (AS199) bound to Staphylococcus A cells (Pansorbin, Calbiochem, La Jolla, CA). The mixtures were precipitated at 8,000 × g for 2 min at 4 °C, and the precipitates were washed 4 times with RIPA buffer (see above). The immunoprecipitates were dissolved in reducing sample buffer and analyzed by SDS-PAGE. The gel was fixed with sodium acetate, impregnated with 15% (w/v) sodium salicylate for 30 min at room temperature, dried for 2 h at 55 °C, and exposed to a Fuji X-Ray film for 24 h at 80 °C.

Effect of Domain-specific Antibodies on BK Liberation-- Freshly drawn human citrated plasma (25 µl) was preincubated with 67.5 µg/ml (450 nM) affinity-purified antibodies to PPK apple domains (A1 to A4) or its light chain, and to HK domain D6_H for 15 min at 37 °C, followed by the addition of FXIIa to a final concentration of 18 µg/ml (20 nM). Following incubation for 1 h, the proteins were separated by reducing SDS-PAGE, and the integrity of the plasma HK, probed by Western blotting using antibodies MBK3 to the kinin portion (D4), was determined.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Our strategy to identify relevant interaction site(s) of the PPK heavy chain with the corresponding acceptor site on the HK light chain comprised (i) the analysis of various kallikrein forms generated by limited proteolysis, (ii) the study of HK binding to recombinantly expressed single apple domains of PPK, (iii) competition studies with antibodies interfering with PPK·HK complex formation, (iv) the construction and HK binding analysis of PPK mutants where single apple domains had been deleted, (v) the construction and binding analysis of PPK-FXI chimeras where relevant apple domains had been exchanged, and (v) the study of effects of antibodies directed to the identified interaction sites of PPK and HK on FXIIa-mediated kinin liberation.

Binding of HK to Various Kallikrein Forms-- Initially the binding of HK to three forms of kallikrein, namely PPK, -kallikrein, and -kallikrein was determined. The extent of conversion of PPK to -kallikrein by FXIIa, and the autocatalytic conversion of -kallikrein to -kallikrein, was confirmed by SDS-PAGE and Western blot analyses using PKH1 and PKH19 antibodies, specific for the A4 and A1 domains, respectively (data not shown). The binding affinity of the three forms of kallikrein for HK was assessed in direct binding assay, in which increasing concentrations of PPK, -kallikrein, and -kallikrein were immobilized on a microtiter plate, and exposed to 8.3 nM biotinylated HK. Bound HK was detected by the streptavidin-peroxidase system. As shown in Fig. 1, biotinylated HK bound with high affinity to PPK (apparent K_D = 6 nM) and -kallikrein (4 nM), whereas the affinity to -kallikrein was drastically reduced (90 nM). Since -kallikrein differs from -kallikrein in having a single peptide bond cleavage at Lys¹⁴⁰ in the A2 domain (9), this result indicates that the integrity of A2 is important to the high affinity binding of HK.

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Cleavage of PPK apple domain A2 results in loss of HK binding capacity. Microtiter plates were coated with plasma prekallikrein (), -kallikrein (), or -kallikrein (), followed by incubation of serial dilutions (2n; starting concentration 100 nM = 8.8 µg/ml) followed by biotinylated HK (8.3 nM = 1 µg/ml). The kallikrein-bound H-kininogen was probed by the streptavidin-peroxidase system. The set-up of the assay is depicted on the upper left; the shaded portion denotes the titer plate, the filled triangle represents the coating protein, and the open box with an asterisk the biotinylated probe.

HK Binding to Single PPK Apple Domains-- To further investigate the role of PPK apple domain A2 for HK binding we cloned and recombinantly expressed single apple domains as fusion proteins with tPA in eukaryotic BHK cells (Fig. 2, left panel), and as fusion proteins with MBP in prokaryotic E. coli cells (Fig. 2, right panel). The purity of the affinity-purified fusion proteins was monitored by SDS-PAGE under reducing conditions. The recombinantly expressed proteins were essentially homogenous except for the tPA-A4 fusion where minor degradation products were observed (Fig. 2A). The various apple fusions with tPA or MBP, the carrier proteins tPA and MBP, and -kallikrein (control; not shown) were coated on microtiter plates and their binding capacity probed by unmodified HK. Bound HK was detected by a monoclonal antibody HKH14 specific for the HK heavy chain not involved in complex formation between HK and PPK. HK bound with highest affinity to immobilized apple domain A2 (set 100%) irrespective of the fusion partner or expression system used (Fig. 2B). The relative binding capacity to other apple domains was reduced to 39-40 (A1) and 48-52% (A4) whereas HK binding to A3 was essentially identical to background binding to the carrier proteins, tPA and MBP (Fig. 2B). These findings are consistent with the notion that A2 plays a pivotal role in the binding of HK to PPK.

View larger version (34K): [in this window] [in a new window]   Fig. 2.   Single apple domains bind differentially to HK. Single PPK apple domains were recombinantly expressed in BHK cells as fusion proteins with tPA (left) or E. coli fused to MBP (right). A, following purification, 100 ng (1.8 pmol) of domains A1, A2, A3, A4, and unfused protein were separated by SDS-PAGE under reducing conditions and visualized by silver staining technique. B, microtiter plates were coated with 100 nM (5 µg/ml) of recombinant constructs, followed by incubation with HK (33 nM = 4 µg/ml). The complex formation between HK and the fusion proteins was probed by HKH14 antibody directed to the heavy chain of HK and a horseradish peroxidase-coupled secondary antibody, followed by the chromogenic substrate. Mean ± S.D. from three independent experiments are presented. The set-up of the assay is given on the right; the first antibody is marked "Y," and the second horseradish peroxidase-labeled antibody identified by an asterisk. Others symbols are as in Fig. 1.

Interference of Antibody PKH6 with HK·PPK Complex Formation-- In a previous study we have developed a panel of 20 monoclonal antibodies to PPK, of which 11 are directed to the PPK heavy chain; they map to four distinct epitope classes arbitrarily designated A-D (22). Taking advantage of the collection of single apple domains which cover the entire PPK heavy chain, we sought to correlate epitope classes A-D with domains A1-A4. Western blot analyses demonstrated that epitopes A, B, and D are localized on apple A4, while epitope C recognized by antibody PKH6 is present on apple A2 (Fig. 3A, exemplified for PKH1 of epitope class A and PKH6, respectively). In another study (15) we have produced monoclonal antibody PKH19 against synthetic peptide PK31; this antibody recognized apple A1 (Fig. 3A, upper line). Since none of the available antibodies bound to apple A3, we used the recombinant tPA-A3 fusion protein of this study to raise a specific antibody to A3 (dubbed -rA3; Fig. 3A); this antibody readily cross-reacts with native PPK (not shown).

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Domain-specific antibodies interfere with PPK·HK complex formation. A, 20 nM each of the MBP-apple domains were separated by SDS-PAGE and probed by Western blotting using antibodies PKH1, PKH6, PKH19, and -rA3 to PPK. B, microtiter plates coated with 4 µg/ml (45.4 nM) PPK were incubated with 1 µg/ml (8.3 nM) biotinylated HK and serial dilutions (2n; starting concentration 180 µg/ml = 1.2 µM) of antibodies PKH19 (), PKH6 (), -rA3 (), PKL16 (), and PKH1 (). Bound HK was detected by the streptavidin-peroxidase system. A representative result of three independent experiments is shown. The setup of the assay is outlined on the top; the symbols as described in the legends of Figs. 1 and 2 are used.

Using the panel of antibodies to each of the four apple domains of PPK, we measured the relative potency of each antibody to interfere with PPK·HK complex formation. To this end we set up a competitive ELISA in which native PPK was bound to a microtiter plate, and serial dilutions of antibodies PKH19 (A1), PKH6 (A2), -rA3 (A3), PKH1 (A4), and PKL16 (directed to the PPK light chain) in the presence of 1 µg/ml biotinylated HK were added. Of the five antibodies tested PKH6 efficiently blocked HK binding to PPK with an apparent IC₅₀ of 8 nM (Fig. 3B). Antibodies PKH19 to A1 and PKH1 to A4 inhibited with IC₅₀ values of 1 µM and 40 nM, respectively, whereas -rA3 and PKL16 (IC₅₀ > 3 µM) failed to interfere with HK·PPK complex formation. Hence A1, A2, and A4 seem to mediate HK docking to PPK although to different extents, whereas A3 is likely not involved. Antibody PKH6 is most potent in inhibiting HK·PPK complex formation; this finding is underlined by the previous observation that the loss of the PKH6 epitope, e.g. by limited proteolysis of -kallikrein, is paralleled by a complete loss of the HK binding capacity of kallikrein (22).

HK Binding to PPK Deletion Mutants-- To further analyze the contribution of the single apple domains to HK binding in the context of the PPK molecule, we employed a loss-of-function model. To this end we constructed PPK mutants in which each of the apple domains was deleted by a PCR-based gene excision technique (Fig. 4A). We transiently transfected the constructs into HEK293t cells which do not endogenously express PPK (30). After 72 h culture supernatants were collected and analyzed by SDS-PAGE and Western blotting. Immunoprinting proved functional expression of the various mutants (Fig. 4B): antibody PKH1 directed to apple domain A4 readily recognized the 1, 2, and 3 deletion constructs lacking apples 1, 2, and 3, respectively, but not the 4 mutant devoid of apple 4, while antibody PKH6 directed to A2 recognized the 1, 3, and 4 mutants, but not the 2 construct. Full-length PPK and an unrelated protein served as positive and negative controls, respectively (Fig. 4B). The concentration of the deletion mutant proteins in cell supernatants was determined by sandwich ELISA and biospecific interaction analysis (data not shown). Next, we tested for the HK binding capacity of the various deletion mutants. Recombinant proteins (22 nM) were coated on microtiter plates, followed by incubation with biotinylated HK and the streptavidin-peroxidase detection system. Under these conditions HK bound strongly to recombinant PPK, which was set as 100%, and to the 3 mutant with almost identical affinity (96.3%). The binding of HK to the 1 and 4 mutants was moderately reduced to 79.7 and 72.5%, respectively, while the 2 mutant showed a drastic loss of HK binding capacity to 34.7% (Fig. 4C). The supernatant of cells expressing an irrelevant protein indicated that the nonspecific binding was <20% of the specific binding. These results show that A2 is indispensable for HK binding, and that apple domains A1 and A4 but not A3 may contribute to the corresponding docking site.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Deletion of apple domain A2 reduces PPK binding to HK. A, schematic diagram of the PPK deletion mutants. The  symbol followed by a number identifies the apple domain that has been deleted. The preformed heavy and light chain portions of single-chain PPK are indicated by lines above the constructs. B, HEK293t cells were transiently transfected with pcDNA3(+) vectors encoding wild-type PPK, the indicated deletion mutants, or an unrelated protein for control (cont). The supernatants (25 µl each) were subjected to SDS-PAGE under reducing conditions, followed by Western blotting with antibody PKH1 to apple domain A4 or with antibody PKH6 to A2. C, microtiter plates were coated with 22 nM recombinant proteins, and incubated with 1 µg/ml (8.3 nM) biotinylated HK followed by 1 µg/ml streptavidin-peroxidase complex and the substrates. Relative HK binding capacity of the constructs is defined as the percentage relative to that obtained with wild-type PPK, where the HK binding capacity is assigned a value of 100%. Mean ± S.D. of three independent experiments are shown. Top right, schematic set-up of the assay; symbols are as described in the legend to Fig. 1.

HK Binding to PPK-FXI Chimeras-- Since apple domain A2 proved to be crucial for HK binding in a loss-of-function model, we wondered whether this domain could serve to transfer high-affinity HK binding in a gain-of-function model. To maintain the structural context, we chose FXI the only other human protein known to contain tandem apple domains, and constructed four FXI-PPK chimeras (Fig. 5A) in which the A2 domains (6.2 versus 6.3) or the complete heavy chains (7.1 versus 7.2) had been homologously exchanged between PPK and FXI. HEK293t cells were transiently transfected with the corresponding constructs, and functional expression of the chimeras was monitored by metabolic labeling with [³⁵S]Cys/Met and immunoprecipitation. Polyclonal antibodies to PPK (AS176) and FXI (AS199) were used to immunoprecipitate radiolabeled PPK-FXI chimeras from the cell supernatants. The immunoprecipitates were resolved by SDS-PAGE under reducing conditions and visualized by autoradiography. These studies show that the constructs were expressed at grossly diverging levels (Fig. 5B). Biospecific interaction analysis used to quantitate the recombinant proteins revealed that they were present in concentrations of 1.2 ± 0.15 µg/ml (chimeras 6.2 and 6.3), 0.8 ± 0.1 µg/ml (7.1), 5.2 ± 0.2 µg/ml (7.2), and 0.6 ± 0.1 µg/ml for wild-type FXI. The direct binding assay (see above) demonstrated that wild-type FXI bound biotinylated HK at a level of 56.4% as compared with PPK (100%), confirming the inherent difference in affinity of the two related proteins for HK (Fig. 5C). Selective transfer of PPK apple domain A2 (chimera 6.3) or the entire PPK heavy chain (7.1) to FXI raised the HK binding affinity to 86.9 and 88.2%, respectively, which were close to that of wild-type PPK. On the other hand incorporation of FXI apple domain A2 (6.2) or the complete FXI heavy chain (7.2) into PPK reduced the HK binding capacity of the corresponding chimeras to 49.9 and 54.8%, respectively, of that of wild-type PPK. These data strongly support the hypothesis that A2 in PPK is responsible for the high-affinity binding to HK.

View larger version (26K): [in this window] [in a new window]   Fig. 5.   Transfer of PPK apple domain A2 increases HK binding activity of FXI. A, scheme of the domain structures of PPK (black), FXI (white), and of PPK-FXI chimeras. B, HEK293t cells were transiently transfected with pcDNA3(+) vectors encoding wild-type FXI, PPK-FXI chimeras, or unrelated protein for control (cont). Cells were metabolically labeled with [35S]Cys/Met, and the supernatants precipitated using a mixture of antibodies to PPK and FXI. Immunoprecipitates were resolved by reducing SDS-PAGE and visualized by autoradiography. C, microtiter plates were coated with 0.5 µg/ml (6 nM) wild-type proteins, recombinant chimeras, and an unrelated protein. Following incubation with 1 µg/ml (8.3 nM) biotinylated HK bound was detected by the streptavidin-peroxidase system. The HK binding capacity of the constructs is expressed as the percentage of that of wild-type PPK (100%). Mean ± S.D. of three independent experiments are given.

Effect of Apple-directed Antibodies on BK Release from HK-- Does the structural importance of apple domain A2 for HK binding bear functional implications on the kallikrein-mediated kinin release from HK? To address this question, we tested the effect of apple domain-specific antibodies on FXIIa-induced kinin release in human plasma. Samples of plasma were preincubated with antibodies directed to PPK apple domains A1 (PKH19), A2 (PKH6), A3 (-rA3), and A4 (PKH1), and to the light chain portion (PKL16); the latter antibody does not interfere with the catalytic activity of the enzyme (22). We also included an antibody to HK domain D6_H (HKL16) directed to the PPK-binding site on the HK light chain portion, and an antibody to a neighboring epitope on D6_H (HKL22) which does not overlap the PPK-binding site. Following preincubation with the respective antibodies, FXIIa was added for 1 h, and then the reaction was stopped. The samples were probed for the presence of uncleaved HK using Western blotting with MBK3 antibodies directed to the kinin moiety (Fig. 6). Antibodies interfering with PPK·HK complex formation, i.e. PKH19, PKH6, PKH1, and HKL16 protected HK from cleavage and therefore prevented BK release. We quantified the amount of intact HK in the various samples and found that the protective effect of PKH6 is equivalent to that of antibody HKL16 directed to the PPK-binding site on HK. Antibodies PKH1 to A4 and PKH19 to A1 were less effective, and antibodies -rA3 to A3 and HKL22 to an irrelevant epitope of D6_H failed to attenuate kinin release. Together these results underline the unique role of apple domain A2 for HK binding to PPK, and stress the importance of physical association of zymogen and prohormone for efficient effector release.

View larger version (26K): [in this window] [in a new window]   Fig. 6.   Antibodies to mutual HK/PPK-binding sites interfere with BK liberation. Fresh human citrate plasma samples were preincubated with 450 nM each of antibodies PKH19 (to A1 of PPK), PKH6 (A2), -rA3 (A3), PKH1 (A4), PKL16 (light chain), or antibodies HKL16 and HKL22 (to domain D6H of HK). Following incubation with 20 nM FXIIa for 1 h at 37 °C, 0.1 µl of plasma was separated by SDS-PAGE and analyzed for the presence of uncleaved HK (116 kDa) by Western blotting using MBK3 to the kinin moiety of HK. Inset, relative inhibition of kinin release is defined as the percentage of intact HK in the samples relative to intact HK in native plasma (not shown) which is assigned a value of 100%. The target epitopes of the antibodies are indicated; D6H antibodies: HKL16 (left), HKL22 (right).

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

H-kininogen circulates in plasma in the form of binary complexes with plasma prekallikrein or factor XI. A considerable fraction of plasma HK docks to the surface of cardiovascular cells such as neutrophils, platelets, and endothelial cells where it anchors to acceptor structures of the plasma membrane (31-33). PPK or FXI remain bound to cell-associated HK; in this way the proenzymes are indirectly attached to cardiovascular cells (34). The assembly of zymogens and prohormone on cell surfaces is thought to serve the circumscribed release of the extremely short-lived effector hormone, bradykinin, in juxtaposition to its target receptors exposed in large number on the endothelium lining the vessels (35, 36). BK is a powerful stimulator of vascular permeability, most probably through opening of endothelial tight junctions (37). Therefore the interaction of HK with PPK is of critical importance to the biological role of kinins in the cardiovascular system.

In this study we have employed a molecular biology approach to define more precisely the relative contributions of the various apple domains to the discontinuous HK-binding site of PPK. The results from the various experimental strategies converge at the conclusion that apple domain A2 forms the principal platform to which HK docks (Fig. 7). Flanking domains A1 and A4 assist A2 in creating a surface accommodating HK, and domain A3 is not involved. The prominent role of apple domain A2 in HK binding is supported by several lines of evidence: (i) proteolytic cleavage of a single peptide bond (Lys¹⁴⁰-Ala¹⁴¹) in A2 drastically reduces the affinity of -kallikrein for HK; (ii) among the recombinantly expressed single apple domains, A2 binds HK with the highest affinity; (iii) removal of A2 leads to the most profound impact on the HK binding capacity of apple deletion mutants; (iv) transfer of PPK A2 increases the HK binding activity of the acceptor construct, FXI, almost to the level of wild-type PPK; (v) in a panel of 22 monoclonal antibodies to PPK, antibody PKH6 directed to A2 most effectively displaces HK from PPK. The peculiar features of antibody PKH6 which has been extensively used for PPK-HK interaction analyses (16, 22, 38) are further highlighted by the finding that proteolysis of its target epitope almost nullifies the HK binding capacity of the corresponding cleavage product, -kallikrein (22).

View larger version (16K): [in this window] [in a new window]   Fig. 7.   PPK apple domain A2 provides the major docking site for HK. The model depicts the HK-binding site of PPK to which HK docks via a continuous site present in domain D6H of its light chain portion. Apple domain A2 considered the major platform for HK is flanked by apple domains A1 and A4 which also contribute to the discontinuous binding site.

Our studies with recombinantly expressed single, combined, and chimeric apple domain constructs bear important structural and functional implications for the PPK-HK interaction. First, the finding that single apple domains bind to HK may indicate that each domain serves as a module that adopts correct folding spontaneously and independently of the other domains of the heavy chain. This notion is corroborated by the finding that deletion mutants lacking single apples retain their HK binding capacity, although at a reduced level. Second, binding of HK was observed independently of the source of the expression system, i.e. eukaryotic versus bacterial cells. This finding suggests that glycosylation (39) does not play a critical role in the complex formation between HK and PPK. Third, the combined data of this study indicate that A3 is not accessible to HK in the native conformation of the PPK heavy chain. This finding is well reflected by the observation that among the 22 distinct monoclonal antibodies produced against native PPK (22), none binds to A3. Fourth, transplantation of PPK to the FXI heavy chain raises the HK binding capacity of the resultant chimera almost to the level of wild-type PPK.

One of the limitations of the present approach is that we have strictly focused on intact apple domains, and therefore we cannot entirely exclude the possibility that inter-domain sequences may also contribute to HK binding. Indeed it has been shown that peptide PK251 (Thr²⁵¹- Gly²⁸⁰) which covers the linker region between A3 and A4, and further extends into the HK-binding site of apple domain A4 (Lys²⁶⁶-Gly²⁹⁵) competes with PPK for binding to HK, although with low efficiency (IC₅₀ = 1.4 mM; note that, the K_D for the native complex is 12 nM) (18). On the other hand, structural analyses with whole apple domains which likely have an "intact" disulfide bridge pattern may help avoid problems encountered in studies using linear peptides. For example, previous work (18) has demonstrated that peptide PK143 (Tyr¹⁴³-Ala¹⁷⁶) encompassing the carboxyl-terminal portion of A2 fails to compete with PPK binding to HK (IC₅₀ > 3 mM). This finding may indicate that the tail region of A2 is not involved in HK binding; however, it may also reflect the failure of the peptide PK143 to fold into the "native" conformation required for HK binding. The complexity of this issue is once more highlighted by antibody PKH6: although its target domain has been unequivocally identified, we have been unable to pinpoint the corresponding target sequence(s) by linear peptides or fusion proteins overlapping the entire A2 domain sequence demonstrating that the PKH6 epitope is highly discontinuous.²

Sequence analysis has revealed that PPK and FXI share 58% identity on the protein level, with a similar disulfide bond pattern in the apple domains except that FXI apple 4 has an unpaired Cys residue mediating homodimerization of the FXI heavy chain (11, 40). Although highly similar in its overall structure this "cousin" protease differs from PPK in some important functional aspects. The compound structure of FXI offers multiple target sites for interacting proteins, and Walsh and co-workers (41) have analyzed the interplay between FXI and various substrates and binding proteins in great detail. Their studies indicate that FXI apple domain A1 binds to HK via a sequence segment (41) that mirrors the corresponding HK attachment site of PPK A1 (15). In addition FXI apple A1 binds to thrombin (42), an important endogenous activator of FXI in blood coagulation (43, 44). FXI domain A3 has been shown to bind to platelets (45) and heparin (46), and domain A4 to FXIIa (47). Using a synthetic peptide approach Walsh and co-workers (48) demonstrated that FXI domain A2 binds FIX although this conclusion has been challenged by the recent study of Sun and Gailani (30) who used recombinant FX-PPK chimeras to show that the FIX-binding site is located in A3 rather than in A2. At present it is unclear whether FXI apple domain A2 has a prime role in HK binding. Our data indicate that incorporation of FXI A2 in construct 6.2 (Fig. 5) increases the affinity for HK by a small but significant increment as compared with deletion mutant 2 (Fig. 4) lacking this domain. The recent advent of a FXI RNA splice variant (49) characterized by a deletion within apple A2 may help to address this intriguing question.

Mapping studies defining interaction sites between molecules involved in biological signaling cascades have proven valuable in identifying novel targets for drug intervention at multiple stages. In the case of the kinin-generating system, the precise mapping of critical sites involved in the local assembly of proenzymes, prohormones, adaptors, and possibly modulators on cell surfaces (50, 51) may help to design novel strategies aimed at blocking the formation of the powerful vascular permeability factor, bradykinin, on levels upstream of its cognate receptors.

	FOOTNOTES

^* This work was supported in part by Deutsche Forschungsgemeinschaft Grant Mu 598/4-2, Fonds der Chemischen Industrie Grant 163323, and National Institutes of Health Grant HL16929.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ Established Investigator of The Netherland's Heart Foundation supported by Grant D96.021.

To whom correspondence should be addressed: Institute of Physiological Chemistry and Pathobiochemistry, Johannes Gutenberg University at Mainz, Duesbergweg 6, D-55099 Mainz, Germany. Tel.: 49-6131-395-890; Fax: 49-6131-395-792; E-mail: wme@mail.uni-mainz.de.

² T. Renné, unpublished observations.

	ABBREVIATIONS

The abbreviations used are: PPK, plasma prekallikrein; BHK, baby hamster kidney cells; BK, bradykinin; ELISA, enzyme-linked immunosorbent assay; HEK293t, human embryonic kidney cells; HK, high-molecular mass kininogen; MBP, maltose-binding protein; PCR, polymerase chain reaction; PAGE, polyacrylamide gel electrophoresis; tPA, tissue plasminogen activator; PBS, phosphate-buffered saline; PCR, polymerase chain reaction.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike